News
People, publications, workshops, and developments.
New preprint on a new computational approach for designing diagonal frame decompositions using data.
Joint work with Andrea Ebner and Dirk Lorenz. Thanks for this nice collaboration!
New preprint on non-convex simple bilevel optimization
What’s the computational hardness of solving non-convex simple bilevel problems. Our recent preprint Stochastic Nonconvex Bilevel Optimization: Improved Rates Without Rare-Visit Assumption asks this question. It has been accepted at OPT 2026: Optimization for Machine Learning (NeurIPS 2026 Workshop).
Sisi Qu moves to Hong Kong Polytechnic University
Sisi will join the group of Defeng Sun as a postdoctoral researcher. Congratulations!
New preprint on inertial forward-backward methods
Inertial forward-backward algorithm with exterior penalization and Tikhonov regularization, with Juan Peypouquet and Sisi Qu.
Optimization in the Länd
We are organizing the IFIP TC-7 workshop with the mathematical optimization and scientific computing labs in Mannheim and Heidelberg.